Skip Navigation

InitialsDiceBearhttps://github.com/dicebear/dicebearhttps://creativecommons.org/publicdomain/zero/1.0/„Initials” (https://github.com/dicebear/dicebear) by „DiceBear”, licensed under „CC0 1.0” (https://creativecommons.org/publicdomain/zero/1.0/)V

vapeloki

@ vapeloki @lemmy.world

Posts
21
Comments
605
Joined
3 yr. ago

  • GenAI vs. ML / Know what we are fighting against

    Jump

  • A language model is any model that operates on language in general. They vary in capabilities. Language recognition is one example of a language usecase.

    Large language models is the class of models that are used for ChatGPT and other modern chatbots.

  • GenAI vs. ML / Know what we are fighting against

    Jump

  • Thanks for the explainer.

    Yes I oversimplified a lot, as I think the audience for this post are people that are not technically deep into the topic.

    Maybe a post about something like from LeNet to LLM could be something in between deep knowledge and current hype. Just not sure where

  • GenAI vs. ML / Know what we are fighting against

    Jump

  • Great writeup! Linked it in the post.

  • GenAI vs. ML / Know what we are fighting against

    Jump

  • You hate GenAI

  • GenAI vs. ML / Know what we are fighting against

    Jump

  • Yes, that is confusing.

    LeNet, the oldest neural network I can name out of my head, was designed to recognize ZIP codes. 1990ish. It is the grandmother of ChatGPT. It does not generate new content. It uses an input set and translates it. Same for image classification in general. We take an input and translate it into a bunch of classes.

    An LLM expands it's input. So it generates NEW content.

  • GenAI vs. ML / Know what we are fighting against

    Jump

  • Yes, but the scale is different. For reference. A conversation with an LLM with around 500kb of raw text requires 20gb and more VRAM just to iterate over this conversation. For each new generated word, the whole input must be reprocessed.

  • GenAI vs. ML / Know what we are fighting against

    Jump

  • I did not use the term AGI? Generative is not a typo and the term I know for AI the generates content instead of just acting on it.

    And to create something like AGI, we would need to adopt different ways to train a Model, like World Models. But that would get into deep AI theroy fast. And I would think that would be out of place here.

  • Fuck AI @lemmy.world
    vapeloki @lemmy.world

    GenAI vs. ML / Know what we are fighting against

  • Las Vegas integrating AI technology into 911 call system

    Jump

  • Did you had a look how this works. Spoiler. Only a small part isn an LLM. Most of it is not.

    Not everything is an LLM. If your car has a line hold assist, most likely, to that is AI. Yes, even the 15 year olds.

    AI != AI

  • Las Vegas integrating AI technology into 911 call system

    Jump

  • That seems actually like a rather intelligent use of AI. Translation, location identification.

    We get promised AI will cure cancer and safe the climate. While not that big of a thing, at least this one saves lifes

  • Don’t you lie to me

    Jump

  • Looks like a bunch of worthless paper to me

  • Former UK biometrics watchdog’s new book argues AI surveillance can be deployed responsibly | Biometric Update

    Jump

  • "Person of interest" and 1984 should be works if fiction and not tutorials

  • American frontier AI labs are asking the US government to lead an international framework of regulation

    Jump

  • So, they are asking to give them the power to regulate themselves.

    US laws and politics is, was, and always will be about money, rich, and fuck privacy.

  • bark rule

    Jump

  • PSA: Liking the feeling of barking could be a verbal stim and a sign auf autism.

    WWWWWoff

  • How OpenAI’s human mistake led to the "AI"-powered hack on Hugging Face

    Jump

  • And I never heard such a story before. Does not sound like anthropic marketing

  • Fuck AI @lemmy.world
    vapeloki @lemmy.world

    Sorry, could not resist

  • After Killings, Homan Says There Will Be ‘More Bloodshed’ Unless Dems ‘Shut Their Mouth’ About ICE | Common Dreams

    Jump

  • If I only could remember where I heard something like this already?

    Ah, right, school, German history, that was it.

  • Removed

    Just some good ole’ fashioned painful plop-shit humour with a hint of race

    Jump

  • Racist AI slop. Blocklist. Btw. I see AI slop I block

  • Haven't generated an APK for months inside Android Studio

    Jump

  • 100% Agreement

  • Haven't generated an APK for months inside Android Studio

    Jump

  • I use the German digital id since years. Without enforced Google.

    Because the German (older) app does not depend on it.

    So no, that's not true. It is an assumption and pessimism

  • Haven't generated an APK for months inside Android Studio

    Jump

  • You have to enable it. And I would love a dedicated switch.

    Currently, you need to enable third party install sources.

    And one could argue that the fact that most people will not enable developer mode is a clear sign that this is a protection.

    In my very very personal opinion: people that do not want to enable developer mode should maybe not install apps from third parties. Developer mode itself unlocks a few switches, and none of them is enabled per default. This is a pretty safe action IMHO.

    There is a way to avoid this, but this is the device owner mode. You need an app that takes full control over device ownership, and those apps are allowed to do anything including installing any app.

    One example is OwnDroid.

    An opensource app that is signed by a dev and allows to manage install sources would be a working solution with better security.

    If someone builds it

  • Fuck AI @lemmy.world
    vapeloki @lemmy.world
    from "AI may not be used..." to "AI must not be used"

    icculus update rules on SDL project on generative AI code:

    github.com /libsdl-org/SDL/blob/main/AGENTS.md
  • Fuck AI @lemmy.world
    vapeloki @lemmy.world

    ‘160% data center carbon footprint increase,’ research finds: Microsoft, Big Oil forge $7B partnership as several AI infrastructure projects push hyperscaler off ‘credible path’ toward decarbonization

    stand.earth /press-releases/160-data-center-carbon-footprint-increase-research-finds-microsoft-big-oil-forge-7b-partnership-as-several-ai-infrastructure-projects-push-hyperscaler-off-credible-path/
  • Linux @lemmy.ml
    vapeloki @lemmy.world

    Unified Gaming Gear linux control - an idea

  • Linux Gaming @lemmy.world
    vapeloki @lemmy.world

    Unified Gaming Gear linux control - an idea

  • Autism @lemmy.world
    vapeloki @lemmy.world

    I am lost, I need advice (TW: death, illness)

  • weedtime @crazypeople.online
    vapeloki @lemmy.world

    12 Hours later: still kicking

  • weedtime @crazypeople.online
    vapeloki @lemmy.world

    Rescue Plants, let's hope the best

  • wir_iel @feddit.org
    vapeloki @lemmy.world

    Wir_iel

  • [Dormant] Electric Vehicles (Moved to !electricvehicles@slrpnk.net) @lemmy.world
    vapeloki @lemmy.world

    I am driving a BMW i5 Touring for two month now, how it goes

  • de_EDV @feddit.de
    vapeloki @lemmy.world

    Must-Have Etiketten für ITler (Print file im Text)

  • ich_iel @feddit.de
    vapeloki @lemmy.world

    ich_iel

  • Cannabis @feddit.de
    vapeloki @lemmy.world

    Oh nein!

  • Deutschland @feddit.de
    vapeloki @lemmy.world

    Endlich Bubatz - 15 Jahre Kampf

  • Cannabis @feddit.de
    vapeloki @lemmy.world

    Endlich - Rückblick auf meine letzten 15 Jahre Aktivismus

  • Cannabis @feddit.de
    vapeloki @lemmy.world

    Dreyer rechnet nicht mit Scheitern von Cannabis-Gesetz - WELT

    www.welt.de /regionales/rheinland-pfalz-saarland/article250652187/Dreyer-rechnet-nicht-mit-Scheitern-von-Cannabis-Gesetz.html
  • Memes @lemmy.ml
    vapeloki @lemmy.world

    Lemmy users looking at stock photos

  • aww @lemmy.world
    vapeloki @lemmy.world

    our 14 year old boy

  • Laser Engraving @lemmy.world
    vapeloki @lemmy.world

    One of our customers wanted some donkeys for kids to paint